1 COVID-19 serology at population scale: SARS-CoV-2-specific antibody responses in saliva 2 3 Authors 4 Pranay R. Randad 1* , Nora Pisanic 1* , Kate Kruczynski 1 , Yukari C. Manabe 2,3 , David Thomas 2 , 5 Andrew Pekosz 1,4 , Sabra L. Klein 4,5 , Michael J. Betenbaugh 6 , William A. Clarke 3 , Oliver 6 Laeyendecker 2,7,8 , Patrizio P. Caturegli 3,9 , H. Benjamin Larman 3,9 , Barbara Detrick 3,9 , Jessica K. 7 Fairley 10 , Amy C. Sherman 11 , Nadine Rouphael 11 , Srilatha Edupuganti 11 , Douglas A. Granger 12 , 8 Steve W. Granger 13 , Matthew Collins 11 , Christopher D. Heaney 1,6,14 9 10 Affiliations 11 *Contributed equally 12 1 Department of Environmental Health and Engineering, Bloomberg School of Public Health, 13 Johns Hopkins University, Baltimore, Maryland, USA 14 2 Division of Infectious Diseases, Department of Medicine, School of Medicine, Johns Hopkins 15 University, Baltimore, Maryland, USA 16 3 Department of Pathology, School of Medicine, Johns Hopkins University, Baltimore, Maryland, 17 USA 18 4 Department of Molecular Microbiology and Immunology, Bloomberg School of Public Health, 19 Johns Hopkins University, Baltimore, Maryland, USA 20 5 Department of Biochemistry and Molecular Biology, Bloomberg School of Public Health, Johns 21 Hopkins University, Baltimore, Maryland, USA 22 6 Department of Chemical and Biomolecular Engineering, Whiting School of Engineering, Johns 23 Hopkins University, Baltimore, Maryland, USA 24 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300 doi: medRxiv preprint NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.
43
Embed
COVID-19 serology at population scale: SARS-CoV-2-specific … · 2020-05-24 · Matched serum and saliva SARS-CoV-2 107 antigen-specific IgG responses were significantly correlated.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
COVID-19 serology at population scale: SARS-CoV-2-specific antibody responses in saliva 2
3
Authors 4
Pranay R. Randad1*, Nora Pisanic1*, Kate Kruczynski1, Yukari C. Manabe2,3, David Thomas2, 5
Andrew Pekosz1,4, Sabra L. Klein4,5, Michael J. Betenbaugh6, William A. Clarke3, Oliver 6
Laeyendecker2,7,8, Patrizio P. Caturegli3,9, H. Benjamin Larman3,9, Barbara Detrick3,9, Jessica K. 7
Fairley10, Amy C. Sherman11, Nadine Rouphael11, Srilatha Edupuganti11, Douglas A. Granger12, 8
Steve W. Granger13, Matthew Collins11, Christopher D. Heaney1,6,14 9
10
Affiliations 11
*Contributed equally 12
1Department of Environmental Health and Engineering, Bloomberg School of Public Health, 13
Johns Hopkins University, Baltimore, Maryland, USA 14
2Division of Infectious Diseases, Department of Medicine, School of Medicine, Johns Hopkins 15
University, Baltimore, Maryland, USA 16
3Department of Pathology, School of Medicine, Johns Hopkins University, Baltimore, Maryland, 17
USA 18
4Department of Molecular Microbiology and Immunology, Bloomberg School of Public Health, 19
Johns Hopkins University, Baltimore, Maryland, USA 20
5Department of Biochemistry and Molecular Biology, Bloomberg School of Public Health, Johns 21
Hopkins University, Baltimore, Maryland, USA 22
6Department of Chemical and Biomolecular Engineering, Whiting School of Engineering, Johns 23
Hopkins University, Baltimore, Maryland, USA 24
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
A multiplex immunoassay to detect SARS-CoV-2-specific antibodies in saliva performs with 46
high diagnostic accuracy as early as ten days post-COVID-19 symptom onset. Highly sensitive 47
and specific salivary COVID-19 antibody assays could advance broad immuno-surveillance 48
goals in the USA and globally. 49
50
Author Contributions 51
All authors reviewed and edited all sections of the article. P.R.R. and N.P. wrote the first draft of 52
the manuscript. K.K. and N.P handled laboratory logistics and generated data. N.P. and P.R.R. 53
analyzed and summarized the data. A.P. provided input on study design and edited the 54
manuscript. M.J.B., S.W.G., D.A.G. provided input on antigen selection, assay design, and 55
interpretation of results. Y.C.M and D.T. provided input on study design and interpretation of 56
results. B.D. provided input on interpretation of results. W.A.C, O.L., P.P.C., and B.L. shared 57
samples and data for the analysis and provided input on interpretation of results. M.H.C 58
developed project concept. M.H.C, N.R., J.F., and A.C.S. led and coordinated specimen 59
collection efforts and reviewed and edited the article. C.D.H. developed project concept and 60
guided the laboratory work. 61
62
Conflict of interest 63
In the interest of full disclosure, D.A.G. is founder and Chief Scientific and Strategy Advisor at 64
Salimetrics, LLC and Salivabio, LLC and these relationships are managed by the policies of the 65
committees on conflict of interest at Johns Hopkins School of Medicine and the University of 66
California at Irvine. N.R. received funds from Sanofi Pasteur, Quidel, Merck and Pfizer. 67
68
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
No. IRB00247886) and by the Emory University Institutional Review Board (IRB No. 91
00110683). 92
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
Non-invasive SARS-CoV-2 antibody testing is urgently needed to estimate the incidence 94
and prevalence of SARS-CoV-2 infection at the general population level. Precise knowledge of 95
population immunity could allow government bodies to make informed decisions about how and 96
when to relax stay-at-home directives and to reopen the economy. We hypothesized that salivary 97
antibodies to SARS-CoV-2 could serve as a non-invasive alternative to serological testing for 98
widespread monitoring of SARS-CoV-2 infection throughout the population. We developed a 99
multiplex SARS-CoV-2 antibody immunoassay based on Luminex technology and tested 167 100
saliva and 324 serum samples, including 134 and 118 negative saliva and serum samples, 101
respectively, collected before the COVID-19 pandemic, and 33 saliva and 206 serum samples 102
from participants with RT-PCR-confirmed SARS-CoV-2 infection. We evaluated the correlation 103
of results obtained in saliva vs. serum and determined the sensitivity and specificity for each 104
diagnostic media, stratified by antibody isotype, for detection of SARS-CoV-2 infection based 105
on COVID-19 case designation for all specimens. Matched serum and saliva SARS-CoV-2 106
antigen-specific IgG responses were significantly correlated. Within the 10-plex SARS-CoV-2 107
panel, the salivary anti-nucleocapsid (N) protein IgG response resulted in the highest sensitivity 108
for detecting prior SARS-CoV-2 infection (100% sensitivity at ≥10 days post-SARS-CoV-2 109
symptom onset). The salivary anti-receptor binding domain (RBD) IgG response resulted in 110
100% specificity. Among individuals with SARS-CoV-2 infection confirmed with RT-PCR, the 111
temporal kinetics of IgG, IgA, and IgM in saliva were consistent with those observed in serum. 112
SARS-CoV-2 appears to trigger a humoral immune response resulting in the almost 113
simultaneous rise of IgG, IgM and IgA levels both in serum and in saliva, mirroring responses 114
consistent with the stimulation of existing, cross-reactive B cells. SARS-CoV-2 antibody testing 115
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
in saliva can play a critically important role in large-scale “sero”-surveillance to address key 116
public health priorities and guide policy and decision-making for COVID-19. 117
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
The coronavirus disease 2019 (COVID-19) pandemic, caused by severe acute respiratory 119
syndrome virus 2 (SARS-CoV-2), has caused >5.4 million COVID-19 cases and >344,000 120
deaths, as of May 24, 2020, involving all populated continents.1 The USA accounts for >1.6 121
million COVID-19 cases and >97,000 deaths and the outbreak has expanded from urban to rural 122
areas of the country.1 There is a critical need to perform broad-scale population-based testing to 123
improve COVID-19 prevention and control efforts. Some have even recommended national 124
testing at repeated time points to improve understanding of the spatio-temporal dynamics of 125
transmission, infection, and herd immunity.2,3 Currently, population-level antibody testing is 126
largely performed using blood, with preliminary seroprevalence study estimates ranging from 127
2.8.% in Santa Clara County, California,4 4.65% in Los Angeles County, California,5 21% in 128
New York City,6 11.5% in Robbio Italy,7 and 14% in Gangelt, Germany.8 Achieving such 129
comprehensive national testing goals will be challenging by relying only on traditional blood-130
based diagnostic specimens as these may be considered too invasive, uncomfortable, or 131
unacceptable, particularly among vulnerable and susceptible groups.9-12 132
In addition to molecular COVID-19 diagnostics, accurate serological tests can identify 133
individuals who have mounted an antibody response to SARS-CoV-2 infection. These tests are 134
needed in platforms that can be deployed in large numbers to describe changes in population 135
level immunity at different geographical scales and over time. Such serological testing could 136
guide “back-to-work” risk mitigation strategies2,3, particularly if evidence continues to emerge 137
suggesting that robust SARS-CoV-2 antibody responses might confer protection from repeated 138
infection.13,14 139
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
Saliva harvested from the space between the gums and the teeth is enriched with gingival 140
crevicular fluid (GCF). The composition of GCF (hereafter referred to as “saliva”) resembles that 141
of serum, and is enriched with antibodies.15-24 Thus, sampling saliva with an appropriate 142
collection method is an attractive non-invasive approach for antibody-based diagnostic 143
techniques. We have previously demonstrated the utility of saliva-based serology testing for the 144
diagnosis, surveillance, and study of infection by multiple viral pathogens.21,22 Development of 145
improved antibody assays to detect prior infection with SARS-CoV-2 has been identified as one 146
of the top unmet needs in the ongoing COVID-19 pandemic response.2,3 Precise knowledge of 147
SARS-CoV-2 infection at the individual level can potentially inform clinical decision-making, 148
whereas at the population level, precise knowledge of prior infection, immunity, and attack rates 149
(particularly asymptomatic infection) is needed to prioritize risk management decision-making 150
about social distancing, treatments, and vaccination (once the latter two become available).25 If 151
saliva can support measurements of both the presence of SARS-CoV-2 RNA26-28 as well as 152
antibodies against SARS-CoV-2, this sample type could provide an important opportunity to 153
monitor individual and population-level SARS-CoV-2 transmission, infection, and immunity 154
dynamics over place and time. 155
Prior studies have shown that antibodies to SARS-CoV-2 nucleocapsid protein (N), spike 156
protein (S), and the receptor binding domain (RBD) are elevated in serum around 10-18 days 157
following SARS-CoV-2 infection.14,29-32 Many ELISA, point-of-care (POC), and lateral flow IgG 158
assays for detecting prior SARS-CoV-2 infection that are currently available show a wide range 159
in diagnostic performance. The sensitivity of the assays improves when samples are collected 160
later after the onset of infection, from <20% sensitivity at <5 days to approximately 100% 161
sensitivity at 17 to 20 days from symptom onset.33-35 162
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
In this study, we aimed to determine whether salivary SARS-CoV-2-specific antibody 163
responses would identify prior SARS-CoV-2 infection with similar sensitivity and specificity as 164
serum and whether salivary antibody testing would reflect the temporal profiles observed in 165
serum. The objectives of this study were: (1) to develop and validate a multiplex bead-based 166
immunoassay for detection of SARS-CoV-2-specific IgG, IgA, and IgM responses; (2) to 167
describe the assay performance using saliva compared to using serum specimens; (3) to identify 168
SARS-CoV-2 antigens that could result in high sensitivity and specificity to identify antibody 169
responses to prior SARS-CoV-2 infection; and (4) to compare the antibody kinetics in saliva to 170
those in serum by time since onset of COVID-19 symptoms. 171
172
Methods 173
Sources of saliva and serum 174
Saliva and serum samples were provided by collaborators from Emory University from 175
patients in three settings: 1) PCR-confirmed COVID-19 cases while admitted to the hospital; 2) 176
confirmed COVID-19 cases we invited to donate specimens after recovering from their acute 177
illness; and 3) patients with symptoms consistent with COVID-19 being tested at an ambulatory 178
testing center donated specimens at the time of testing and/or at a follow-up convalescent phase 179
research visit. Collaborators at Johns Hopkins University provided: 1) serum samples from 180
patients presenting with COVID-19-like symptoms such as fever, cough, dyspnea who were 181
recruited in both inpatient and outpatient clinical cares sites; and 2) negative saliva and serum 182
samples collected prior to the COVID-19 pandemic. Participants provided verbal and / or written 183
informed consent and provided saliva and blood specimens for analysis. Whenever possible, 184
remnant clinical blood specimens were used. Basic data on days since symptom onset were 185
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
recorded for all participants as were results of COVID-19 molecular testing. Participation in 186
these studies was voluntary and the study protocols have been approved by the respective 187
Institutional Review Boards. 188
189
Saliva and blood sample collection 190
Saliva samples were collected by instructing participants to gently brush their gum line 191
with an Oracol S14 saliva collection device (Malvern Medical Developments, UK) for 1-2 192
minutes, or until saturation. This saliva collection method specifically harvests GCF, which is 193
enriched with primarily IgG antibody derived from serum.18 The saturated sponge was then 194
inserted into the storage tube, capped, and stored at 4○C until processing whenever possible. 195
Saliva was separated from the Oracol S14 swabs through centrifugation (10 min at 1,500 g) and 196
transferred into the attached 2 mL cryovial. Samples were heat-inactivated at 60○C for 30 197
minutes and then shipped to the lab on dry ice. Blood samples were collected into ACD (acid, 198
citrate, dextrose) or serum separator tubes (SST) and processed according to each clinical lab’s 199
procedure. Plasma/serum was also heat inactivated at 60○C for 30 minutes, aliquoted into 2mL 200
cryovials, and stored at ≤20°C until analyzed. Only de-identified serum or plasma and saliva 201
aliquots including limited metadata (days since symptom onset and SARS-CoV-2 RT-PCR status 202
[ever positive or negative]) were shared for this study. 203
204
Multiplex magnetic microparticle (“bead”)-based SARS-CoV-2 saliva immunoassay 205
Ten SARS-CoV-2 antigens were obtained commercially or from collaborators at Icahn 206
School of Medicine at Mount Sinai (Table 1).36 This included four SARS-CoV-2 receptor 207
binding domain (RBD), one ectodomain (ECD) protein containing the S1 and S2 subunit of the 208
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
spike protein, two S1 subunits, one S2 subunit, and two N proteins. Each SARS-CoV-2 antigen, 209
along with one SARS-CoV-1 antigen (NAC SARS 2002 N) and one human coronavirus (hCoV)-210
229E antigen (Sino Biol. hCoV 229E ECD), were covalently coupled to magnetic microparticles 211
(MagPlex microspheres, Luminex) as described previously (Table 1).21,22 Along with a control 212
bead, conjugated with bovine serum albumin (BSA), the multiplex panel included a total of 13 213
bead sets (10 bead sets coupled to SARS-CoV-2 antigens, one to SARS-CoV-1 antigen, one to 214
hCoV-229E antigen, and one control bead coupled to BSA). Coupling of antigens to beads was 215
confirmed using antibody against the antigen or against the tag (e.g. anti-His(6) tag antibody), if 216
present (Table 1), followed by a species-specific R-phycoerythrin (PE)-labelled antibody and 217
was considered successful if the median fluorescence intensity (MFI [a.u.]) was >10,000 at 1 218
μg/mL of antigen-specific antibody (except the BSA-conjugated bead set). Saliva samples were 219
centrifuged (5 minutes at 20,000g, 20°C), and 10 μL of saliva supernatant was added to 40 μL of 220
assay buffer (phosphate-buffered saline with 0.05% Tween20, 0.02% sodium azide and 1% 221
BSA) containing 1,500 beads of each bead set per microplate well. The plate was covered and 222
incubated at room temperature for 1 hour on a plate shaker at 500 rpm. Beads were washed twice 223
with 200 μL PBST and 50 μL of PE-labeled anti-human IgG, IgA or IgM diluted 1:100 in assay 224
buffer were added, and the plate was incubated again for 1 hour on a plate shaker at 500 rpm. 225
Beads were washed as above and then suspended in 100 μL of assay buffer. Finally, the MFI of 226
each bead set was measured on a Bio-Plex® immunoassay instrument (Bio-Rad Laboratories, 227
Hercules, CA). The same protocol was used for serum and plasma samples, except that serum 228
and plasma samples were tested at a final dilution of 1:1000 in bead mix and assay buffer 229
compared to a final dilution of 1:5 for saliva. A subset of 47 saliva samples were tested in 230
duplicate and in a masked fashion to determine intra-assay variability (same 96 well plate) and 231
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
inter-assay variability (different 96 well plates on different days), and at least 2 blanks (assay 232
buffer) were included on each plate for background fluorescence subtraction. 233
234
Statistical analysis 235
The median fluorescence intensity (MFI) measured using the BSA beads was subtracted 236
from each blank-subtracted antigen-specific MFI signal for each sample to account for non-237
specific binding of antibodies to beads. The average MFI was used for samples that were tested 238
in duplicate (n=47) or triplicate. Wilcoxon-Mann-Whitney test was used to compare the median 239
MFI between samples collected <10 days post symptom onset and negatives, and between 240
samples collected ≥10 days post symptom onset and negatives, for each antigen in the multiplex. 241
The average intra- and inter-assay variability was evaluated by determining the coefficient of 242
variation (CV%) of a subset of 47 samples that were tested in duplicate (intra) and on different 243
days and plates (inter). Pearson’s correlation was used to determine the correlation between 244
antigen-specific IgG, IgA, and IgM MFI in matched saliva and serum / plasma samples collected 245
from the same person at the same time point (n=28). The average MFI of all saliva samples from 246
known uninfected individuals (pre-Covid-19) plus three standard deviations for each antigen-247
specific IgG, IgA, and IgM were used to establish the cut-off values for a negative result. The 248
corresponding procedure was used for serum samples. Because the prior hCoV infection status 249
for saliva and serum samples was not known, the MFI cut-off values were not calculated for anti-250
Sino Biol. hCoV 229E ECD IgG, IgA, and IgM. Sensitivity and specificity for detecting samples 251
from confirmed RT-PCR positive individuals and for samples from individuals obtained prior to 252
the COVID-19 pandemic were determined for each antigen/isotype pair (IgG, IgM and IgA) in 253
saliva and in serum. Locally weighted regression (LOESS) was used to visualize and compare 254
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
Information on days post symptom onset was collected for each positive participant. A total of 262
134 saliva samples (from 2012 to early 2019) and 112 serum samples (from 2016)37 were 263
collected from participants enrolled in cohort studies prior to the start of the COVID-19 264
pandemic and were designated as negative samples (pre-COVID-19 pandemic) (Table 2). 265
266
SARS-CoV-2 antigen-specific IgG, IgA, and IgM cut-off values 267
The multiplex immunoassay, comprised of ten SARS-CoV-2 antigens (2 N proteins, 1 268
ECD protein, four RBD proteins, two S1 subunits, and one S2 subunit), one SARS-CoV-1 269
antigen (NAC SARS CoV 2002 N), and one hCoV-229E antigen (Sino Biol. hCoV 229E ECD) 270
was used to test a total of 167 saliva samples from 150 individuals and 324 serum samples from 271
171 individuals. The range, median, mean, standard deviation, and derived MFI cut off value for 272
each saliva and serum SARS-CoV-2 antigen-specific IgG, IgA, and IgM stratified by negative 273
samples, samples collected <10 days, and ≥10 days post SARS-CoV-2 symptom onset are 274
provided in Supplementary Table 1 and Supplementary Table 2. Saliva collected at ≥10 days 275
post symptom onset had significantly elevated IgG levels (median MFI) against all SARS-CoV-2 276
antigens compared to negative saliva samples (Supplementary Table 1). Serum collected at ≥10 277
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
days post symptom onset had significantly elevated IgG, IgA, and IgM levels (median MFI) 278
against all SARS-CoV-2 antigens compared to negative sera. 279
280
Correlation between saliva and serum SARS-CoV-2-specific IgG 281
Twenty-eight participants provided matched saliva and serum samples that were collected 282
during the same visit (n=6 negative and n=22 RT-PCR confirmed SARS-CoV-2 infection 283
matched saliva and serum samples). Antigen-specific IgG levels in matched saliva and serum 284
samples were significantly correlated for all SARS-CoV-2 and SARS-CoV-1 antigens (Figure 285
1). Antigen-specific IgA in matched saliva and serum samples were modestly correlated with 286
significance detected only for a subset of antigens: GenScript N, Sino Biol. N, Sino Biol. ECD, 287
GenScript S1, and NAC SARS 2002 N (Figure 2). Antigen-specific IgM in matched saliva and 288
serum samples were also significantly correlated for all SARS-CoV-2 and SARS-CoV-1 289
antigens, although the correlation was weaker than for IgG (Figure 3). 290
291
Saliva: Sensitivity and specificity 292
In saliva, the sensitivity to detect SARS-CoV-2 infection increased among saliva samples 293
collected ≥10 days post symptom onset compared to those collected <10 days post symptom 294
onset, for all isotypes (IgG, IgA, and IgM)(Figure 4). The highest sensitivity (100%) was 295
achieved with GenScript N-coupled beads in saliva samples collected ≥10 days post symptom 296
onset. All (28/28) individuals with RT-PCR confirmed prior SARS-CoV-2 infection had salivary 297
anti-GenScript N IgG levels above the cut-off (Figure 4). Specificity to classify negative saliva 298
samples correctly ranged from 98% to 100% for SARS-CoV-2 IgG. Mt. Sinai’s RBD resulted in 299
the highest specificity (100%). All (134/134) negative saliva samples resulted in MFI values 300
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
below the cut-off (mean + 3 SD) for anti-Mt. Sinai RBD IgG levels. The highest combined 301
sensitivity and specificity was achieved with GenScript N (100% sensitivity and 99% specificity 302
at ≥10 days post symptom onset). 303
While IgA and IgM against SARS-CoV-2 also remained equivalent or increased among 304
saliva samples collected ≥10 days compared to <10 days post symptom onset, the sensitivity to 305
detect prior SARS-CoV-2 infection remained low (Figure 4). For SARS-CoV-2 specific IgA, 306
sensitivity ranged from 4% with NAC S2 to 61% with Sino Biol. ECD. For IgM, sensitivity 307
ranged from 0% with NAC S2 to 65% with GenScript S1. Specificity for IgA ranged from 42% 308
with GenScript S1 to 100% with NAC S1 and S2. The highest combined sensitivity and 309
specificity for IgA was obtained with Sino Biol. ECD (61% sensitivity ≥10 days post symptom 310
onset and 96% specificity). For IgM, specificity ranged from 96% (GenScript RBD [i]) to 99% 311
(Sino Biol. ECD, GenScript S1, and NAC S2). The highest combined sensitivity and specificity 312
for IgM was reached with GenScript S1 (65% sensitivity, 99% specificity). 313
314
Serum: Sensitivity and specificity 315
In serum, the sensitivity to detect SARS-CoV-2 infection improved among serum 316
samples collected ≥10 days compared to <10 days post symptom onset, for all isotypes (IgG, 317
IgA, and IgM)(Figure 5). For anti-SARS-CoV-2 IgG, the highest sensitivity (92%) achieved 318
with Mt. Sinai and Sino Biol. RBD using sera collected ≥10 days post symptom onset (96/104 319
samples from individuals with RT-PCR confirmed prior SARS-CoV-2 infection had IgG levels 320
against these antigens above the cut-offs) (Figure 5). Specificity ranged from 96%-99% for anti-321
SARS-CoV-2 IgG. The highest combined sensitivity and specificity was achieved with Mt. 322
Sinai’s RBD (92% sensitivity and 99% specificity at ≥10 days post symptom onset) 323
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
For anti-SARS-CoV-2 IgA and IgM, sensitivity ranged from 0% to 45% when using 324
serum samples collected <10 days post-COVID-19 symptom onset; sensitivity was higher 325
overall when detecting IgA compared to IgM. The sensitivity improved significantly with several 326
antigens (predominantly RBD), when samples collected ≥10 days post-COVID-19 symptom 327
onset were tested (Figure 5). When testing these sera, the highest sensitivity to detect IgA was 328
reached using GenScript RBD (h) antigen (95%; 99/104 samples above the cutoff) but several 329
additional antigens also performed with high sensitivity. In contrast, only two antigens (Mt. Sinai 330
RBD and GenScript RBD [h]) in the assay reached sensitivities above 90% when detecting anti-331
SARS-CoV-2 IgM. Specificity ranged from 96%-99% for both anti-SARS-CoV-2 IgA and IgM. 332
The highest combined sensitivity and specificity for detecting IgA and IgM was reached using 333
Mt. Sinai’s RBD (as was the case for serum IgG) but also when using NAC’s SARS 2002 N 334
antigen (Figure 5). 335
336
Temporal kinetics of SARS-CoV-2 specific IgG, IgA, and IgM responses in serum compared to 337
saliva 338
The temporal kinetics of antigen-specific IgG, IgA, and IgM responses in serum and in 339
saliva are shown in Figure 6. Also shown are the cut-offs for each isotype (IgG, IgA, and IgM) 340
in serum and in saliva (dashed lines). The temporal kinetics and magnitude of the antigen-341
specific IgG and IgA responses in saliva generally correlate with those detected in serum. The 342
IgM response is significantly lower in magnitude (MFIs) in saliva compared to serum, which is 343
expected and consistent with the lower relative concentration of total IgM in saliva compared to 344
total IgA and IgG concentrations in saliva. 345
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
In serum, the SARS-CoV-specific IgA levels across individuals consistently cross the 346
cut-off (dashed lines), thus indicating seroconversion, several days before IgG and IgM. IgG and 347
IgM seroconversion in serum seem to occur approximately at the same time. 348
Even though saliva IgA levels increase closely after IgG levels, the SARS-CoV-2-349
specific IgA response often does not cross the cut-off, indicative of the low observed sensitivity. 350
IgM levels in saliva are low and LOESS regression lines generally remain under the cut-off for 351
most antigens in the multiplex assay. However, in saliva, the antigen-specific IgG response 352
consistently crosses the cut-off around 10 days post symptom onset, i.e. approximately 15 days 353
post infection, similar, to the time of IgG seroconversion in serum. The anti-SARS-CoV-2 IgG 354
response in saliva thus appears to mimic seroconversion in serum. 355
356
Reactivity of antibodies with SARS-CoV-1 and hCoV proteins following SARS-CoV-2 infection 357
We sought to evaluate reactivity of SARS-CoV-1 and hCoV proteins in samples from 358
COVID-19 cases. For IgG, all convalescent phase saliva from COVID-19 cases (28/28; 100%) 359
reacted with the NAC SARS 2002 N protein. Similarly, 89% and 95% of convalescent sera from 360
COVID-19 cases reacted with the NAC SARS 2002 N protein for IgG and IgA, respectively. 361
The median MFI for salivary IgG and IgA, and serum IgG, IgA, and IgM, to NAC SARS 2002 N 362
was significantly elevated among samples collected ≥10 days post symptom onset compared to 363
negatives (Supplementary Table 1 and Supplementary Table 2). The median MFI for saliva 364
and serum IgG and IgA to Sino Biol. hCoV 229E ECD was also elevated among samples 365
collected ≥10 days post symptom onset compared to negatives (Supplementary Table 1 and 366
Supplementary Table 2). These results suggest that SARS-CoV-2 elicits cross-reactive 367
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
antibodies to the closely related SARS-CoV-1, and that reactivity to Sino Biol. hCoV 229E ECD 368
is very common in our study population, likely due to frequent human exposure to hCoVs. 369
370
Intra- and inter-assay variability 371
Among 47 saliva samples assayed in duplicate on the same 96-well plate, the average 372
intra-assay variability ranged from 3%-18% (CV%) (Supplementary Table 3). Among 47 saliva 373
samples tested in duplicate on different 96-well plates on different days, the average inter-assay 374
variability ranged from 5%-28% (CV%) (Supplementary Table 3). 375
376
Discussion 377
Our results demonstrate that salivary SARS-CoV-2-specific IgG detection reflects the 378
binding profile observed in serum. Salivary SARS-CoV-2-specific IgG can be used to detect a 379
prior SARS-CoV-2 infection with high sensitivity and specificity. When saliva was collected 380
≥10 days post symptom onset, the anti-SARS-CoV-2 IgG assay detects SARS-CoV-2 infection 381
with 100% sensitivity and 99% specificity (GenScript N) and/or with 89% sensitivity and 100% 382
specificity (Mt. Sinai RBD). In addition, we demonstrate that the temporal kinetics of SARS-383
CoV-2-specific IgG responses in saliva are consistent with those observed in serum and indicate 384
that most individuals seroconvert approximately 10 days after COVID-19 symptom onset or 385
approximately two weeks post-presumed infection. Based on these results it is feasible to 386
accurately measure the salivary IgG response to identify individuals with a prior SARS-CoV-2 387
infection. Our saliva-based multiplex immunoassay could serve as a non-invasive approach for 388
accurate and large-scale SARS-CoV-2 “sero”-surveillance. Because saliva samples can be self-389
collected and mailed at ambient temperatures,24 a saliva antibody test could greatly increase the 390
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
scale of testing—particularly among susceptible populations—compared to blood, and could 391
clarify population immunity and susceptibility to SARS-CoV-2. 392
Matched saliva and serum samples demonstrate a significant correlation in SARS-CoV-2 393
antigen-specific IgG responses. An analysis of temporal kinetics of antibody responses in saliva 394
following COVID-19 symptom onset revealed a congruence with those observed in serum, and a 395
synchronous elevation of SARS-CoV-2 serum IgG and IgM responses, which has been reported 396
in serum.14,29-32 In both saliva and serum, IgG rather than IgM was the first isotype to increase, 397
mimicking a response consistent with the stimulation of existing, cross-reactive B cells, even 398
though this is a novel coronavirus in these human populations. Both synchronous and classical 399
antibody isotype responses have been previously reported following SARS-CoV-2 infection.14,29-400
32 Furthermore, IgG levels in saliva and serum tended to rise and cross the cut-off around day 10 401
post-COVID-19 symptoms onset, which is typically when individuals seek care from a 402
healthcare provider for the first time. Therefore, salivary antibody testing could be used in 403
combination with standard SARS-CoV-2 nucleic acid diagnostic testing to provide critical 404
information about antibody positivity and temporal kinetics, which can be informative for patient 405
trajectories and outcomes. 406
The sensitivity of our assay improved or remained the same among saliva and serum 407
samples collected during convalescent phase (≥10 days post symptom onset) compared to acute 408
phase (<10 days post symptom onset) for all SARS-CoV-2 antigen-specific IgG, IgA, and IgM. 409
Saliva SARS-CoV-2 antigen-specific IgG peaked at 100% sensitivity, and serum SARS-CoV-2 410
antigen-specific IgG at 92% sensitivity (anti-Sino Biol. RBD IgG and anti-Mt. Sinai RBD IgG, 411
respectively) among samples collected ≥10 days post SARS-CoV-2 symptom onset. Earlier 412
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
studies have reported sensitivities for various SARS-CoV-2 IgG tests peaking at 82%-100% 413
sensitivity among samples collected during convalescent phase of infection.33-35 414
While serum IgA and IgM peaked at 95% and 93% sensitivity, respectively, at >=10 days 415
post symptom onset, saliva IgA and IgM reached a sensitivity of only 61% and 65%, 416
respectively. The median MFI for most SARS-CoV-2 antigen-specific IgA and IgM responses in 417
saliva were, however, significantly elevated at ≥10 days post symptom onset compared to 418
negative control samples (Supplementary Table 1). One explanation for the low sensitivity 419
observed in saliva for IgA and IgM may be due to the background signal-to-noise ratio for saliva 420
SARS-CoV-2 antigen-specific IgA and IgM, which was greater than that observed for saliva 421
IgG. Non-specific binding of salivary proteins, exogenous particles, non-specific antibodies, or 422
cross-reactivity with other viruses could contribute to this background. Although we harvested 423
GCF, which is enriched with blood transudate, because of size exclusion IgM antibodies are not 424
abundant in saliva (12). Nevertheless, SARS-CoV-2 antigen-specific IgG responses in saliva 425
performed with improved sensitivity and specificity compared to serum, peaking at 100% 426
sensitivity ≥10 days post symptom onset for anti-GenScript N IgG and 100% specificity for anti-427
Mt. Sinai RBD IgG. 428
Virus infections often induce antibody responses that cross-react with related viruses, 429
which can compromise the performance of serologic assays. Cross-reactivity may largely be 430
attributable to the N protein and S2 subunit, which share 90% sequence homology with SARS-431
CoV-1.31 The RBD of the S protein is less conserved across beta-CoVs than the N protein and 432
whole S protein, and many antibodies known to interact with SARS-CoV-1’s RBD do not 433
interact with SARS-CoV-2’s RBD.38 For these reasons, we hypothesized that SARS-CoV-2 N 434
would be highly sensitive and cross-react with antibodies following SARS-CoV-1 infection, 435
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
whereas those against SARS-CoV-2 RBD would be more specific.36 We found that all (28/28; 436
100%) saliva samples from COVID-19 cases collected at ≥10 days post-symptom onset reacted 437
with NAC SARS 2002 N in the IgG assay, indicating that SARS-CoV-2 infection can elicit 438
cross-reactive IgG to closely related CoVs. Of course, this antigen could still be used for SARS-439
CoV-2 diagnostics, as the cross-reactivity would only be relevant if SARS-CoV-1 and SARS-440
CoV-2 were co-circulating in the same human population. We did not specifically evaluate 441
whether common hCoVs elicit cross-reactive antibodies that could cause false positive results in 442
our SARS-CoV-2 assay; however, we did include one hCoV antigen (hCoV-229E ECD) in the 443
panel. Sera from early and late COVID-19 cases and negative control samples all reacted 444
similarly to this antigen, which is consistent with a high prevalence of hCoV exposure in the 445
general population.39-41 This also strongly suggests that our negative control sample population 446
was highly exposed to hCoV and we would not have been able to achieve such clear 447
discrimination between negative control and COVID-19 samples with other antigens in the 448
multiplex panel if cross-reactivity was a significant issue. 449
This study has several limitations. First, our collection of saliva and serum samples was 450
predominantly obtained from independent cohorts, and it contained 28 matched saliva and serum 451
samples collected from the same participants at the same time. In future studies, the performance 452
of this assay should be compared between saliva and serum in a large sample of matched saliva 453
and serum samples. Second, all saliva data was cross sectional and we were not able to evaluate 454
the temporal kinetics of saliva SARS-CoV-2 antibody responses using repeated measures within 455
the same individual. Longitudinal analysis would allow us to evaluate the temporal kinetics and 456
magnitude of SARS-CoV-2 IgG, IgA, and IgM responses, resolve synchronous vs. classical 457
isotype responses (IgM followed by IgA followed by IgG) following SARS-CoV-2 infection.42 458
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
Additional investigation with convalescent phase saliva and sera are needed to determine the 459
stability of SARS-CoV-2-specific IgG responses. Third, we did not have information on severity 460
of SARS-CoV-2 disease from each participant in this study, and thus were not able to determine 461
the impact of severity of infection on antibody responses.29 Prior studies suggest that antibody 462
responses are slightly elevated among individuals with severe infection.29,30,42 Future analysis 463
should determine how severity of infection, and infectious dose, modifies antibody responses. 464
Fourth, we did not determine receiver operating characteristic (ROC)-optimized MFI cut offs in 465
this analysis. However, the cut offs used in this study (average of negatives + three standard 466
deviations) are conservative. Future analysis should identify ROC-optimized cut offs, which 467
could improve the sensitivity and specificity of this saliva assay. Lastly, we did not have 468
sociodemographic and medical history information for participants, and thus were not able to 469
evaluate the relationship of age, sex, and other factors on antibody responses. 470
In future analysis, additional replicates should be used to assess intra- and inter-assay 471
variability, and a lower limit of detection should be determined for each antigen. Furthermore, 472
well characterized sera from other hCoV and zoonotic CoV infections should be used to address 473
potential cross-reactivity of antibodies following SARS-CoV-1, MERS-CoV, hCoV-OC43, 474
hCoV-HKU1, hCoV-229E, and hCoV-NL63 infection with SARS-CoV-2 proteins. Lastly, the 475
performance of this saliva assay should be compared head-to-head with other clinically utilized 476
antibody tests. 477
Saliva represents a practical, non-invasive alternative to NP, OP, blood, and stool-based 478
diagnostic specimens for COVID-19 diagnostic testing. Recently, saliva collection via passive 479
drool (instructing patients to spit into a sterile urine specimen collection cup) was shown to be 480
more sensitive than NP specimens for SARS-CoV-2 RNA detection by RT-PCR in COVID-19 481
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
patients.26 Furthermore, the U.S. Food and Drug Administration recently granted emergency use 482
authorization for a saliva-based nucleic acid test for SARS-CoV-2 that can be collected at home 483
and mailed in for testing.43 Recognition of the advantages of saliva both for SARS-CoV-2 484
nucleic acid and antibody testing could accelerate goals for nationwide testing to surveil active 485
and prior SARS-CoV-2 infections at the general population level. 486
This study demonstrates that SARS-CoV-2 antigen-specific antibody responses in saliva 487
reflect those observed in serum, and that SARS-CoV-2 antigen-specific IgG can be used to 488
accurately detect prior SARS-CoV-2 infection. We have developed and validated a saliva-based 489
multiplex immunoassay and identified SARS-CoV-2 antigen-specific IgG responses that can 490
detect prior SARS-CoV-2 infection with high sensitivity (anti-N IgG; 100% sensitivity, 99% 491
specificity) and specificity (anti-RBD IgG; 89% sensitivity, 100% specificity) at ≥10 days post 492
symptom onset. An accurate saliva-based antibody test for prior SARS-CoV-2 infection would 493
greatly improve our ability to perform public health interventions in the current pandemic. This 494
non-invasive method for comprehensive determination of prior SARS-CoV-2 infection will 495
facilitate large-scale “sero”-surveillance to evaluate population immunity. As SARS-CoV-2 496
vaccine candidates progress through clinical trials, such non-invasive tests will be critical to 497
identify immunity gaps and susceptible populations to inform targeted vaccination efforts, as 498
well as companion diagnostics for vaccine trials.44 Furthermore, saliva assays can be used to 499
monitor correlates of protection and the force of transmission in community-based settings, pre- 500
and post- vaccination/prevention strategies, to determine the effectiveness of population-based 501
interventions and direct future preventative strategies. 502
503
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
1. Dong, E., Du, H. & Gardner, L. An interactive web-based dashboard to track COVID-19 505
in real time. The Lancet Infectious Diseases 20, 533-534 (2020). 506
2. Angulo, F.J., Finelli, L. & Swerdlow, D.L. Reopening Society and the Need for Real-507
Time Assessment of COVID-19 at the Community Level. JAMA (2020). 508
3. Gronvall, G., et al. Developing a National Strategy for Serology ( Antibody Testing ) in 509
the United States. Johns Hopkins - Bloomberg School of Public Health (2020). 510
4. Bendavid, E., et al. COVID-19 Antibody Seroprevalence in Santa Clara County, 511
California. medRxiv, 2020.2004.2014.20062463-512
20062020.20062404.20062414.20062463 (2020). 513
5. Sood, N., et al. Seroprevalence of SARS-CoV-2-Specific Antibodies Among Adults in 514
Los Angeles County, California, on April 10-11, 2020. JAMA (2020). 515
6. Cuomo Says 21% of Those Tested in N.Y.C. Had Virus Antibodies. (2020). 516
7. Zorzoli, M. ESCLUSIVA I nuovi dati di Robbio, unico paese italiano a fare il test 517
sull’immunità a tutti i cittadini. 70% di asintomatici. (2020). 518
8. Regalado, A. Blood tests show 14% of people are now immune to covid-19 in one town 519
in Germany. (2020). 520
9. Dyal, J.W., et al. COVID-19 Among Workers in Meat and Poultry Processing Facilities - 521
19 States, April 2020. MMWR. Morbidity and mortality weekly report 69(2020). 522
10. Lloyd-Sherlock, P., Ebrahim, S., Geffen, L. & McKee, M. Bearing the brunt of covid-19: 523
older people in low and middle income countries. BMJ 368, m1052 (2020). 524
11. Ward, C.F., Figiel, G.S. & McDonald, W.M. Altered Mental Status as a Novel Initial 525
Clinical Presentation for COVID-19 Infection in the Elderly. The American journal of 526
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
26. Wyllie, A.L., et al. Saliva is more sensitive for SARS-CoV-2 detection in COVID-19 563
patients than nasopharyngeal swabs. medRxiv, 2020.2004.2016.20067835-564
20062020.20067804.20067816.20067835 (2020). 565
27. Sullivan, P.S., et al. Detection of SARS-CoV-2 RNA and Antibodies in Diverse Samples: 566
Protocol to Validate the Sufficiency of Provider-Observed, Home-Collected Blood, 567
Saliva, and Oropharyngeal Samples. JMIR public health and surveillance 6, e19054 568
(2020). 569
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
37. Kelen, G.D., et al. Improvements in the continuum of HIV care in an inner-city 590
emergency department. AIDS 30, 113-120 (2016). 591
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
41. Gorse, G.J., Patel, G.B., Vitale, J.N. & O'Connor, T.Z. Prevalence of antibodies to four 601
human coronaviruses is lower in nasal secretions than in serum. Clinical and vaccine 602
immunology : CVI 17, 1875-1880 (2010). 603
42. Long, Q.-X., et al. Antibody responses to SARS-CoV-2 in patients with COVID-19. 604
Nature Medicine (2020). 605
43. FDA. ACCELERATED EMERGENCY USE AUTHORIZATION (EUA) SUMMARY 606
SARS-CoV-2 ASSAY (Rutgers Clinical Genomics Laboratory). Vol. 2020 (US Food and Drug 607
Administration, 2020). 608
44. Lipsitch, M., Kahn, R. & Mina, M.J. Antibody testing will enhance the power and 609
accuracy of COVID-19-prevention trials. Nature Medicine (2020). 610
611
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
N his anti-Gen N GenScript N Z03480 A02039 Sino Biol. N his anti-Sino N Sino Biol. N 40588-V08B A02039 Sino Biol. ECD (S1+S2) his anti-Sino RBD Sino Biol. ECD (S1+S2) 40589-V08B1 A02038 Sino Biol. RBD his anti-Sino RBD Sino Biol. RBD 40592-V08H 40592-T62 Mt. Sinai RBD his anti-Sino RBD Mt. Sinai RBD Amanat F., et al 40592-T62
GenScript RBD his anti-Sino RBD Sino Biol. RBD (h) Z03479 40592-T62 GenScript RBD his anti-Sino RBD Sino Biol. RBD (i) Z03483 40592-T62 GenScript S1 N/A anti-Gen S GenScript S1 Z03501 A02038
NAC S1 shFc anti-Sheep Fc NAC S1 REC31806 313-005-046 NAC S2 shFc anti-Sheep Fc NAC S2 REC31807 313-005-046 NAC SARS-CoV-1 SARS CoV N his anti-his NAC SARS 2002 N REC31744 MA121315
Sino Biol. hCoV-229E 229E ECD his anti-his Sino Biol. hCoV 229E ECD 40605-V08B MA121315 *Sino Biol.: Sino Biological; NAC: Native Antigen Company #N: nucleocapsid protein; ECD: ectodomain (S1 + S2 subunit of spike protein); RBD: receptor binding domain
^Corresponding IgG antibody used for confirmation
613
614
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
Figure 1. Correlation between saliva and serum SARS-CoV-2 antigen-specific IgG among 632
matched saliva and serum samples (n=28). Pearson correlation coefficient is provided for each 633
antigen-specific IgG. p values are provided for statistically significant correlations only (p<0.05). 634
Note. Sino Biol.: Sino Biological; NAC: Native Antigen Company; N: nucleocapsid protein; 635
ECD: S1: S1 subunit of spike protein; S2: S2 subunit of spike protein; ectodomain (S1 636
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
subunit+S2 subunit of the spike protein); RBD: receptor binding domain; (h): produced in human 637
cell; (i): produced in insect cell; MFI=mean fluorescence intensity. 638
639
640
641
642
643
644
645
646
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
Figure 2. Correlation between saliva and serum SARS-CoV-2 antigen-specific IgA among 648
matched saliva and serum samples (n=26). Pearson correlation coefficient is provided for each 649
antigen-specific IgA. p values are provided for statistically significant correlations only (p<0.05). 650
Note. Sino Biol.: Sino Biological; NAC: Native Antigen Company; N: nucleocapsid protein; 651
ECD: S1: S1 subunit of spike protein; S2: S2 subunit of spike protein; ectodomain (S1 652
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
subunit+S2 subunit of the spike protein); RBD: receptor binding domain; (h): produced in human 653
cell; (i): produced in insect cell; MFI=mean fluorescence intensity. 654
655
656
657
658
659
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
Figure 3. Correlation between saliva and serum SARS-CoV-2 antigen-specific IgM among 661
matched saliva and serum samples (n=26). Pearson correlation coefficient is provided for each 662
antigen-specific IgM. p values are provided for statistically significant correlations only 663
(p<0.05). Note. Sino Biol.: Sino Biological; NAC: Native Antigen Company; N: nucleocapsid 664
protein; ECD: S1: S1 subunit of spike protein; S2: S2 subunit of spike protein; ectodomain (S1 665
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
subunit+S2 subunit of the spike protein); RBD: receptor binding domain; (h): produced in human 666
cell; (i): produced in insect cell; MFI=mean fluorescence intensity. 667
668
669
670
671
672
673
674
675
676
677
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
Figure 4. The sensitivity and specificity of each SARS-CoV-2 antigen-specific IgG, IgA, and 680
IgM in saliva. Samples collected from individuals with RT-PCR confirmed prior SARS-CoV-2 681
infection are stratified into samples collected <10 days post symptom onset and samples 682
collected ≥10 days post symptom onset. The average MFI of negative samples + 3 standard 683
deviations was used to set the MFI cut off for each SARS-CoV-2 antigen-specific IgG, IgA, and 684
IgM. Darker shades of green indicate higher whereas darker shades of red indicate lower 685
sensitivity and specificity. Note. Sino Biol.: Sino Biological; NAC: Native Antigen Company; N: 686
nucleocapsid protein; ECD: S1: S1 subunit of spike protein; S2: S2 subunit of spike protein; 687
ectodomain (S1 subunit+S2 subunit of the spike protein); RBD: receptor binding domain; (h): 688
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
produced in human cell; (i): produced in insect cell; Se: Sensitivity; Sp: specificity; MFI=mean 689
fluorescence intensity. 690
691
692
693
694
695
696
697
698
699
700
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
Figure 5. The sensitivity and specificity of each SARS-CoV-2 antigen-specific IgG, IgA, and 703
IgM in serum. Samples collected from individuals with RT-PCR confirmed prior SARS-CoV-2 704
infection are stratified into samples collected <10 days post symptom onset and samples 705
collected ≥10 days post symptom onset. The average MFI of negative samples + 3 standard 706
deviations was used to set the MFI cut off for each SARS-CoV-2 antigen-specific IgG, IgA, and 707
IgM. Darker shades of green indicate higher whereas darker shades of red indicate lower 708
sensitivity and specificity. Note. Sino Biol.: Sino Biological; NAC: Native Antigen Company; N: 709
nucleocapsid protein; ECD: S1: S1 subunit of spike protein; S2: S2 subunit of spike protein; 710
ectodomain (S1 subunit+S2 subunit of the spike protein); RBD: receptor binding domain; (h): 711
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
produced in human cell; (i): produced in insect cell; Se: Sensitivity; Sp: specificity; MFI=mean 712
fluorescence intensity. 713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
Figure 6. Comparison of saliva and serum SARS-CoV-2 antigen-specific IgG (red), IgA (blue), 735
and IgM (green) responses vs. days post-COVID-19 symptom onset. The trajectories of IgG 736
(red), IgA (blue), and IgM (green) responses are estimated using a LOESS curve. Dashed lines 737
indicate cut off values for IgG (red), IgA (blue), and IgM (green). Note. Sino Biol.: Sino 738
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint
spike protein; S2: S2 subunit of spike protein; ectodomain (S1 subunit+S2 subunit of the spike 740
protein); RBD: receptor binding domain; (h): produced in human cell; (i): produced in insect 741
cell; MFI=mean fluorescence intensity. 742
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 26, 2020. ; https://doi.org/10.1101/2020.05.24.20112300doi: medRxiv preprint